Polysemy resolution with comprehension of emotion detection is a decisive aspect of linguistic processing of Bengali text. The purpose of emotion detection in natural language processing is to increase the accuracy of sentiment and emotion analysis by resolving multiple meanings of words within a given context. In emotion identification, the same word can carry different emotional nuances based on its context, making polysemy resolution essential for precise interpretation. Emotion detection from textual content plays an important role in understanding human communication and behaviour in the digital age. In this context, this paper presents ten polysemous words with three types of emotions, namely, positive, negative and neutral, where each type of emotion consists of thirty documents. Additionally, this paper proposes an innovative concept for polysemy-based feature collection with a combination of measuring the significance of words by considering both their frequency in a document and their rarity across the entire corpus. It employs a resource-scarce artificial neural network (RS-ANN) classifier for revealing emotion from Bengali text with exposure to polysemy resolution.